Detection of formation layer boundaries on well log measurements is required for interpretation, modeling and inversion tasks. Existing methods for automatic formation boundary detection, which are mostly based on data variance and frequency analysis, are not robust and effective. Conventional formation layer boundary detection on well log measurements therefore, is often manually performed and time consuming. These disadvantages are compounded in situations involving multi-sensor data and multi-well data.